stdlib: Generalize erl_parse:abstract/2

The 'encoding' option of erl_parse:abstract/2 has been extended to
include 'none' and a callback function (a predicate).

The rationale is that a more general means of determining what integer
lists are to be represented as strings may help readability when
generating Erlang code given input in some other encoding than Latin-1
or UTF-8.

3 files changed, 65 insertions, 35 deletions

diff --git a/lib/stdlib/doc/src/erl_parse.xml b/lib/stdlib/doc/src/erl_parse.xml
index 2d5aff3c6c..cf0bff48cd 100644
--- a/lib/stdlib/doc/src/erl_parse.xml
+++ b/lib/stdlib/doc/src/erl_parse.xml
@@ -4,7 +4,7 @@
 <erlref>
   <header>
     <copyright>
-      <year>1996</year><year>2013</year>
+      <year>1996</year><year>2014</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -173,6 +173,7 @@
     </func>
     <func>
       <name name="abstract" arity="2"/>
+      <type name="encoding_func"/>
       <fsummary>Convert an Erlang term into an abstract form</fsummary>
       <desc>
         <p>Converts the Erlang data structure <c><anno>Data</anno></c> into an
@@ -183,7 +184,12 @@
           selecting which integer lists will be considered
           as strings. The default is to use the encoding returned by
           <seealso marker="epp#default_encoding/0">
-          <c>epp:default_encoding/0</c></seealso></p>
+          <c>epp:default_encoding/0</c></seealso>.
+          The value <c>none</c> means that no integer lists will be
+          considered as strings. The <c>encoding_func()</c> will be
+          called with one integer of a list at a time, and if it
+          returns <c>true</c> for every integer the list will be
+          considered a string.</p>
       </desc>
     </func>
   </funcs>
diff --git a/lib/stdlib/src/erl_parse.yrl b/lib/stdlib/src/erl_parse.yrl
index 6316db7054..1dc5fc52a7 100644
--- a/lib/stdlib/src/erl_parse.yrl
+++ b/lib/stdlib/src/erl_parse.yrl
@@ -2,7 +2,7 @@
 %%
 %% %CopyrightBegin%
 %%
-%% Copyright Ericsson AB 1996-2013. All Rights Reserved.
+%% Copyright Ericsson AB 1996-2014. All Rights Reserved.
 %%
 %% The contents of this file are subject to the Erlang Public License,
 %% Version 1.1, (the "License"); you may not use this file except in
@@ -919,59 +919,63 @@ normalise_list([]) ->
       Data :: term(),
       AbsTerm :: abstract_expr().
 abstract(T) ->
-    abstract(T, 0, epp:default_encoding()).
+    abstract(T, 0, enc_func(epp:default_encoding())).
+
+-type encoding_func() :: fun((non_neg_integer()) -> boolean()).
 
 %%% abstract/2 takes line and encoding options
 -spec abstract(Data, Options) -> AbsTerm when
       Data :: term(),
       Options :: Line | [Option],
       Option :: {line, Line} | {encoding, Encoding},
-      Encoding :: latin1 | unicode | utf8,
+      Encoding :: 'latin1' | 'unicode' | 'utf8' | 'none' | encoding_func(),
       Line :: erl_scan:line(),
       AbsTerm :: abstract_expr().
 
 abstract(T, Line) when is_integer(Line) ->
-    abstract(T, Line, epp:default_encoding());
+    abstract(T, Line, enc_func(epp:default_encoding()));
 abstract(T, Options) when is_list(Options) ->
     Line = proplists:get_value(line, Options, 0),
     Encoding = proplists:get_value(encoding, Options,epp:default_encoding()),
-    abstract(T, Line, Encoding).
+    EncFunc = enc_func(Encoding),
+    abstract(T, Line, EncFunc).
 
 -define(UNICODE(C),
-        is_integer(C) andalso
-         (C >= 0 andalso C < 16#D800 orelse
+         (C < 16#D800 orelse
           C > 16#DFFF andalso C < 16#FFFE orelse
           C > 16#FFFF andalso C =< 16#10FFFF)).
 
+enc_func(latin1) -> fun(C) -> C < 256 end;
+enc_func(unicode) -> fun(C) -> ?UNICODE(C) end;
+enc_func(utf8) -> fun(C) -> ?UNICODE(C) end;
+enc_func(none) -> none;
+enc_func(Fun) when is_function(Fun, 1) -> Fun;
+enc_func(Term) -> erlang:error({badarg, Term}).
+
 abstract(T, L, _E) when is_integer(T) -> {integer,L,T};
 abstract(T, L, _E) when is_float(T) -> {float,L,T};
 abstract(T, L, _E) when is_atom(T) -> {atom,L,T};
 abstract([], L, _E) -> {nil,L};
 abstract(B, L, _E) when is_bitstring(B) ->
     {bin, L, [abstract_byte(Byte, L) || Byte <- bitstring_to_list(B)]};
-abstract([C|T], L, unicode=E) when ?UNICODE(C) ->
-    abstract_unicode_string(T, [C], L, E);
-abstract([C|T], L, utf8=E) when ?UNICODE(C) ->
-    abstract_unicode_string(T, [C], L, E);
-abstract([C|T], L, latin1=E) when is_integer(C), 0 =< C, C < 256 ->
-    abstract_string(T, [C], L, E);
-abstract([H|T], L, E) ->
+abstract([H|T], L, none=E) ->
     {cons,L,abstract(H, L, E),abstract(T, L, E)};
+abstract(List, L, E) when is_list(List) ->
+    abstract_list(List, [], L, E);
 abstract(Tuple, L, E) when is_tuple(Tuple) ->
-    {tuple,L,abstract_list(tuple_to_list(Tuple), L, E)}.
-
-abstract_string([C|T], String, L, E) when is_integer(C), 0 =< C, C < 256 ->
-    abstract_string(T, [C|String], L, E);
-abstract_string([], String, L, _E) ->
-    {string, L, lists:reverse(String)};
-abstract_string(T, String, L, E) ->
-    not_string(String, abstract(T, L, E), L, E).
-
-abstract_unicode_string([C|T], String, L, E) when ?UNICODE(C) ->
-    abstract_unicode_string(T, [C|String], L, E);
-abstract_unicode_string([], String, L, _E) ->
+    {tuple,L,abstract_tuple_list(tuple_to_list(Tuple), L, E)}.
+
+abstract_list([H|T], String, L, E) ->
+    case is_integer(H) andalso H >= 0 andalso E(H) of
+        true ->
+            abstract_list(T, [H|String], L, E);
+        false ->
+            AbstrList = {cons,L,abstract(H, L, E),abstract(T, L, E)},
+            not_string(String, AbstrList, L, E)
+    end;
+abstract_list([], String, L, _E) ->
     {string, L, lists:reverse(String)};
-abstract_unicode_string(T, String, L, E) ->
+abstract_list(T, String, L, E) ->
     not_string(String, abstract(T, L, E), L, E).
 
 not_string([C|T], Result, L, E) ->
@@ -979,9 +983,9 @@ not_string([C|T], Result, L, E) ->
 not_string([], Result, _L, _E) ->
     Result.
 
-abstract_list([H|T], L, E) ->
-    [abstract(H, L, E)|abstract_list(T, L, E)];
-abstract_list([], _L, _E) ->
+abstract_tuple_list([H|T], L, E) ->
+    [abstract(H, L, E)|abstract_tuple_list(T, L, E)];
+abstract_tuple_list([], _L, _E) ->
     [].
 
 abstract_byte(Byte, L) when is_integer(Byte) ->
diff --git a/lib/stdlib/test/erl_scan_SUITE.erl b/lib/stdlib/test/erl_scan_SUITE.erl
index 447e159cd4..35067e8116 100644
--- a/lib/stdlib/test/erl_scan_SUITE.erl
+++ b/lib/stdlib/test/erl_scan_SUITE.erl
@@ -1,7 +1,7 @@
 %%
 %% %CopyrightBegin%
 %%
-%% Copyright Ericsson AB 1998-2013. All Rights Reserved.
+%% Copyright Ericsson AB 1998-2014. All Rights Reserved.
 %%
 %% The contents of this file are subject to the Erlang Public License,
 %% Version 1.1, (the "License"); you may not use this file except in
@@ -21,7 +21,7 @@
 	 init_per_group/2,end_per_group/2]).
 
 -export([ error_1/1, error_2/1, iso88591/1, otp_7810/1, otp_10302/1,
-          otp_10990/1, otp_10992/1]).
+          otp_10990/1, otp_10992/1, otp_11807/1]).
 
 -import(lists, [nth/2,flatten/1]).
 -import(io_lib, [print/1]).
@@ -60,7 +60,8 @@ end_per_testcase(_Case, Config) ->
 suite() -> [{ct_hooks,[ts_install_cth]}].
 
 all() -> 
-    [{group, error}, iso88591, otp_7810, otp_10302, otp_10990, otp_10992].
+    [{group, error}, iso88591, otp_7810, otp_10302, otp_10990, otp_10992,
+     otp_11807].
 
 groups() -> 
     [{error, [], [error_1, error_2]}].
@@ -1144,6 +1145,25 @@ otp_10992(Config) when is_list(Config) ->
         erl_parse:abstract([$A,42.0], [{encoding,utf8}]),
     ok.
 
+otp_11807(doc) ->
+    "OTP-11807. Generalize erl_parse:abstract/2.";
+otp_11807(suite) ->
+    [];
+otp_11807(Config) when is_list(Config) ->
+    {cons,0,{integer,0,97},{cons,0,{integer,0,98},{nil,0}}} =
+        erl_parse:abstract("ab", [{encoding,none}]),
+    {cons,0,{integer,0,-1},{nil,0}} =
+        erl_parse:abstract([-1], [{encoding,latin1}]),
+    ASCII = fun(I) -> I >= 0 andalso I < 128 end,
+    {string,0,"xyz"} = erl_parse:abstract("xyz", [{encoding,ASCII}]),
+    {cons,0,{integer,0,228},{nil,0}} =
+        erl_parse:abstract([228], [{encoding,ASCII}]),
+    {cons,0,{integer,0,97},{atom,0,a}} =
+        erl_parse:abstract("a"++a, [{encoding,latin1}]),
+    {'EXIT', {{badarg,bad},_}} = % minor backward incompatibility
+         (catch erl_parse:abstract("string", [{encoding,bad}])),
+   ok.
+
 test_string(String, Expected) ->
     {ok, Expected, _End} = erl_scan:string(String),
     test(String).